Advanced Graph Algorithms Spring 2003
Gabriel Valiente
valiente@lsi.upc.es
Technical University of Catalonia
Technical University of Catalonia
1/152
Gabriel Valiente
Schedule
1. Introduction
2. Intersection Graph Theory
3. Algorithms on Permutation Graphs
4. Algorithms on Circle Graphs
5. Algorithms on Interval Graphs
6. Algorithms on Chordal Graphs
7. Algorithms on CircularArc Graphs
8. Student Presentations
13 May 2003 27 May 2003
03 June 2003
10 June 2003
28 May 2003 28 May 2003
04 June 2003 04 June 2003 11 June 2003
Technical University of Catalonia
2/152
Gabriel Valiente
Introduction
Technical University of Catalonia
3/152
Gabriel Valiente
Introduction
• Most graph problems of theoretical interest and practical relevance are intractable, even hard to approximate.
• However, NPhard problems become polynomialtime solvable on special graph classes.
• A closer look will be taken in this course at some fundamental graph problems (clique, independent set, coloring, isomorphism).
• Understanding the structure of the graph classes that make them polynomialtime solvable.
• Drawing the boundary between P and NPhard for them.
• Studying robust (certifying) algorithms.
• Identifying open problems.
• Graph classes will be studied from the point of view of intersection graph theory.
Technical University of Catalonia
4/152
Gabriel Valiente
Introduction
Isomorphism expresses what is meant when two graphs are said to be the same graph.
Deﬁnition
bijection h : V ( G ) → V ( H ) such that, for every pair of vertices u, v ∈ V ( G ) ,
Two graphs G and H are isomorphic, denoted by G ^{∼}
_{=} H, if there is a
{u, v} ∈ E ( G ) if and only if {h ( u ) , h ( v ) } ∈ E ( H ) .
Two isomorphic graphs may be depicted in such a way that they look very different.
Example
The following two graphs are isomorphic.
Technical University of Catalonia
5/152
Gabriel Valiente
Introduction
Nonisomorphism of graphs is not usually hard to prove, because several invariants or necessary conditions for isomorphism are not difﬁcult to compute. These are properties that do not depend on the presentation or labeling of a graph.
Deﬁnition
graphs to be isomorphic.
Example
degree sequence.
Remark
V ( H ) = {v _{1} ,
An graph isomorphism invariant is a necessary condition for two
Two graphs cannot be isomorphic if they differ in their order, size, or
For input graphs G and H with V ( G ) = {u _{1} ,
,v
_{n} }, a necessary condition for G ^{∼}
_{n} } and
,u
_{=} H is that the multisets
{Γ ( u _{i} )  1 i n} and {Γ ( v _{i} )  1 i n} be equal.
Technical University of Catalonia
6/152
Gabriel Valiente
Introduction
Invariants are not sufﬁcient conditions for isomorphism. Nothing can be concluded about two graphs which share an invariant.
Example
same number of vertices, the same number of edges, and are both 4regular.
The following two graphs are not isomorphic, although they have the
Technical University of Catalonia
7/152
Gabriel Valiente
Introduction
Example
same number of vertices, the same number of edges, and are both 4regular.
{u _{0} ,u _{1} },{u _{0} ,u _{2} },{u _{0} ,u _{4} },{u _{0} ,u _{6} },{u _{1} ,u _{3} },{u _{1} ,u _{5} },{u _{1} ,u _{7} },
The following two graphs are not isomorphic, although they have the
{u 
_{3} },{u _{9} },{u 


{u 
_{2} ,u _{5} ,u 
_{2} ,u _{5} ,u 
_{8} },{u _{1}_{3} },{u _{2} ,u _{1}_{0} _{6} ,u _{8} 
},{u _{3} ,u _{4} },{u _{3} ,u _{6} },{u _{4} ,u _{5} },{u _{4} ,u _{7} }, },{u _{1}_{0} },{u _{7} ,u _{9} },{u _{7} ,u _{1}_{3} }, _{6} ,u 
_{1}_{1} },{u _{8}
_{1}_{1} ,u _{1}_{2} },{u _{1}_{1} ,u _{1}_{5} },{u _{1}_{2} ,u _{1}_{3}
{u
{u
_{8} ,u 
,u _{1}_{4} 


},{u _{1}_{2} ,u _{1}_{4} },{u _{1}_{3} ,u _{1}_{5} },{u _{1}_{4}
,u _{1}_{5} }
},{u _{9} ,u _{1}_{2} },{u _{9} ,u _{1}_{5} },{u _{1}_{0} ,u _{1}_{1} },{u _{1}_{0} ,u _{1}_{4} },
E ( G ) =
E ( H ) =
{v _{2} ,v _{3} },{v _{2} ,v _{8} },{v _{2} ,v _{1}_{0} },{v _{3} ,v _{9} },{v _{3} ,v _{1}_{1} },{v _{4} ,v _{5} },{v _{4} ,v _{8} },
{v _{0} ,v _{1} },{v _{0} ,v _{2} },{v _{0} ,v _{4} },{v _{0} ,v _{6} },{v _{1} ,v _{3} },{v _{1} ,v _{5} },{v _{1} ,v _{7} },
{v
_{4} ,v
_{7} ,v
{v
_{1}_{2} },{v
_{1}_{1} },{v
_{5} ,v
_{7} ,v
_{9} },{v _{5} ,v _{1}_{3} },{v
_{1}_{3} },{v _{8} ,v _{9} },{v
_{6} ,v
_{8} ,v
_{7} },{v _{6} ,v _{1}_{0} },{v _{6} ,v _{1}_{2}
},
_{1}_{4} },{v _{9} ,v _{1}_{5} },{v _{1}_{0} ,v _{1}_{1} },
{v
_{1}_{0} ,v _{1}_{4} },{v _{1}_{1} ,v _{1}_{5} },{v _{1}_{2} ,v _{1}_{3} },{v _{1}_{2} ,v _{1}_{4} },{v _{1}_{3} ,v _{1}_{5} },{v _{1}_{4} ,v _{1}_{5} }
Technical University of Catalonia
8/152
Gabriel Valiente
Introduction
Example
same number of vertices, the same number of edges, and are both 4regular.
The following two graphs are not isomorphic, although they have the
There are, in fact, only ﬁve vertices at distance 2 of any given vertex of the graph to the lefthand side, but six vertices at distance 2 of any given vertex of the graph to the righthand side.
Technical University of Catalonia
9/152
Gabriel Valiente
Introduction
Isomorphism of graphs is usually much harder to prove than nonisomorphism of graphs, because all known certiﬁcates or necessary and sufﬁcient conditions for graph isomorphism are as difﬁcult to compute as graph isomorphism itself.
Deﬁnition
condition for two graphs to be isomorphic.
From a complexitytheoretical point of view, graph isomorphism is one of the few NP problems believed neither to be in P nor to be NPcomplete.
An graph isomorphism certiﬁcate is a necessary and sufﬁcient
NP
NPC
·
_{·}
_{·}
P
GI
Technical University of Catalonia
10/152
Gabriel Valiente
Introduction
Deﬁnition
h : V ( G ) → V ( H ) such that, for every pair of vertices u, v ∈ V ( G ) , {h ( u ) , h ( v ) } ∈ E ( H ) if {u, v} ∈ E ( G ) . In an induced subgraph isomorphism, {u, v} ∈ E ( G ) if and only if {h ( u ) , h ( v ) } ∈ E ( H ) .
Deﬁnition
graphs G and H, whether H has a subgraph which is isomorphic to G.
Remark
graph problems.
A subgraph isomorphism of a graph G into a graph H is an injection
The subgraph isomorphism problem is to determine, given two input
Subgraph isomorphism is a common generalization of many important
• Clique (K _{n} )
• Independent set (nK _{1} )
• Hamiltonian cycle (C _{n} )
• Matching (nK _{2} )
• Girth (P _{n} )
• Shortest path (P _{n} )
Technical University of Catalonia
11/152
Gabriel Valiente
Introduction
Deﬁnition
a mapping h : V ( G ) → V ( H ) such that, for every pair of vertices u, v ∈ V ( G ) ,
Two graphs G and H are homomorphic, denoted by G H, if there is
{h ( u ) , h ( v ) } ∈ E ( H ) if {u, v} ∈ E ( G ) .
Deﬁnition
input graphs G and H, whether H has a subgraph which is homomorphic to G.
Remark
adjacent vertices obtain adjacent colors.
Example
The subgraph homomorphism problem is to determine, given two
Graph homomorphism is a generalization of graph coloring, in which
G K _{n} if and only if G is ncolorable.
Technical University of Catalonia
12/152
Gabriel Valiente
Introduction
Deﬁnition
H by deleting degreetwo vertices.
Deﬁnition
input graphs G and H, whether H has a subgraph which is homeomorphic to G.
Remark
graphs characterized by the absence of forbidden structures.
Example (Kuratowski, 1930)
of a subdivision of either K _{5} or K _{3}_{,}_{3} .
A graph G is homeomorphic to a graph H if G can be obtained from
The subgraph homeomorphism problem is to determine, given two
Subgraph homeomorphism is the recognition problem for classes of
Planar graphs are characterized by the absence
Technical University of Catalonia
13/152
Gabriel Valiente
Introduction
Example
subdivision of K _{4} .
Seriesparallel graphs are characterized by the absence of a
Technical University of Catalonia
14/152
Gabriel Valiente
Introduction
Deﬁnition
subgraph of H by contracting edges.
Deﬁnition
graphs G and H, whether H (has a subgraph which) is a minor of G.
Remark
characterized by the absence of forbidden minors.
Example
A graph G is a minor of a graph H if G can be obtained from a
The minor containment problem is to determine, given two input
Minor containment is the recognition problem for classes of graphs
K _{5} is a minor of the Petersen graph.
Technical University of Catalonia
15/152
Gabriel Valiente
graph subgraph common subgraph common supergraph
graph class
Introduction
minor
homeomorphism
homomorphism
isomorphism
Technical University of Catalonia
16/152
Gabriel Valiente
Deﬁnition
Introduction
A certifying algorithm for a decision problem is an algorithm that
provides a certiﬁcate with each answer that it produces.
Example
graph G, a 2coloring of G is an acceptance certiﬁcate, while an odd cycle in G is
a rejection certiﬁcate.
Consider the problem of recognizing bipartite graphs. Given an input
Remark
Certifying algorithms are also called robust.
Technical University of Catalonia
17/152
Gabriel Valiente
Introduction
• Martin C. Golumbic, Algorithmic Graph Theory and Perfect Graphs, New York: Academic Press, 1980.
• Klaus Simon, Efﬁziente Algorithmen für perfekte Graphen, Stuttgart: B. G. Teubner, 1992.
• Johannes Köbler, Uwe Schöning, and Jacobo Turán. The Graph
Isomorphism Problem: its Structural Complexity. Boston: Birkhäuser,
1993.
• Ron Shamir. Advanced Topics in Graph Algorithms. Technical Report, TelAviv University, 1994.
• Terry A. McKee and Fred R. McMorris, Topics in Intersection Graph Theory, Philadelphia: SIAM, 1999.
• Andreas Brandstädt, Van Bang Le, and Jeremy P. Spinrad, Graph Classes:
A Survey, Philadelphia: SIAM, 1999.
• Gabriel Valiente. Algorithms on Trees and Graphs. Berlin:
SpringerVerlag, 2002.
Technical University of Catalonia
18/152
Gabriel Valiente
Intersection Graph Theory
Technical University of Catalonia
19/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
by Ω( F ) , is the graph having F as vertex set with S _{i} adjacent to S _{j} if and only if
i = j and S _{i} ∩S _{j} = 0/ . A graph G is called an intersection graph if there exists a
multiset of sets F such that G ^{∼} _{=} Ω( F ) .
Example
S _{3} = {x _{4} }, and S _{4} = {x _{1} ,x _{3} ,x _{4} ,x _{5} }. Then, G ^{∼}
The intersection graph of a multiset of sets F = {S _{1} ,
,S
_{n} }, denoted
Let F = {S _{1} ,S _{2} ,S _{3} ,S _{4} }, where S _{1} = {x _{1} }, S _{2} = {x _{1} ,x _{2} ,x _{3} },
_{=} Ω( F ) is depicted next.
v
1
v 2
v 4
Theorem (Marczewski, 1945)
v 3
x 1
x _{1} x _{3} x _{4} x _{5}
x 1 x 2 x 3
x _{4}
Every graph is an intersection graph.
Theorem 
Every graph is the intersection graph of a family of subgraphs of a 
graph. 
Technical University of Catalonia
20/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition a tree.
Deﬁnition
along a circle.
Deﬁnition along a line.
Deﬁnition
circle.
Deﬁnition
between an ordered set of vertices and a permutation of them.
A chordal graph is the intersection graph of a ﬁnite set of subtrees of
A circulararc graph is the intersection graph of a ﬁnite set of arcs
An interval graph is the intersection graph of a ﬁnite set of intervals
A circle graph is the intersection graph of a ﬁnite set of chords of a
A permutation graph is the intersection graph of a ﬁnite set of chords
Technical University of Catalonia
21/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
A permutation graph is a graph that is isomorphic to the intersection
graph of a ﬁnite set of chords between an ordered set of vertices and a permutation of them.
Example
The following ﬁnite set of chords between an ordered set of vertices
and a permutation of them is a permutation model or matching diagram of the permutation graph shown to the left.
Remark
e _{6}
e
e
8
1
e 2
e _{7}
e
4
e 5
e
3
v 1 
v 2 
v 3 
v 4 
v 5 
v 6 
v 7 
v 8 

v 3 
v 5 
v 8 
v 4 
v 6 
v 1 
v 7 
v 2 
A permutation graph is the graph of inversions in a permutation.
Technical University of Catalonia
22/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
of a ﬁnite set of chords of a circle.
Example
circle graph shown to the left.
A circle graph is a graph that is isomorphic to the intersection graph
The following ﬁnite set of chords of a circle is a chord model of the
e 3
e 1
e 4
e 5
e 2
e _{6}
e _{7}
v
v
8
v
v
5
v
4
7
9
v 6
v 10
v 3
v
v
v 13
2
v
14
v 11
v 12
1
Technical University of Catalonia
23/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
graph of a ﬁnite set of subtrees of a tree.
Example
chordal graph shown to the left.
A chordal graph is a graph that is isomorphic to the intersection
The following family of subtrees of a tree is a subtree model of the
2 
1 
3456 
3456 

4 
5 

245 
4567 15 
245 4567 15 

3 
7 

6 
24 
467 1 5 
24 
467 1 5 

3456 
3456 
3456 
3456 
3456 

245 4567 15 
245 4567 15 
245 4567 15 
245 
4567 15 
245 4567 15 

24 
467 1 5 
24 
467 1 5 
24 
467 1 5 
24 
467 1 5 
24 
467 1 5 
Technical University of Catalonia
24/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
graph of a ﬁnite set of arcs along a circle.
Example
circulararc graph shown to the left.
A circulararc graph is a graph that is isomorphic to the intersection
The following ﬁnite set of arcs along a circle is an arc model of the
1 

2 

3 
4 

5 
6 
7 
12 
2 

2 
23 

24 
35 

47 
356 

4 
36 

34 
346 
Technical University of Catalonia
25/152
Gabriel Valiente
Intersection Graph Theory
Deﬁnition
graph of a ﬁnite set of intervals along a line.
Example
the interval graph shown to the left.
An interval graph is a graph that is isomorphic to the intersection
The following ﬁnite set of intervals along a line is an interval model of
1
2
3
4
6
5
7
8
1
3
6
8
Technical University of Catalonia
26/152
Gabriel Valiente
Intersection Graph Theory
Lemma
the previous ﬁve graph classes.
The following diagram summarizes the relationships that hold between
circle
permutation
chordal
interval
circulararc
Technical University of Catalonia
27/152
Gabriel Valiente
Intersection Graph Theory
Exercise
between the ﬁve graph classes hold, or give at most 32 examples of graphs that belong to some of the previous ﬁve graph classes.
Example
circulararc graph. Consider, for instance, the intersection graph of three noncrossing chords which all cross two noncrossing chords along the circle graph with equator. This permutation graph is isomorphic to K _{2}_{,}_{3} and has no circulararc model, because two arcs along the circle cannot both intersect three nonintersecting arcs along the circle without intersecting themselves.
Prove the previous lemma. Show that the claimed relationships
A permutation graph is a circle graph, but it need not be a
Technical University of Catalonia
28/152
Gabriel Valiente
Intersection Graph Theory
Example (Hajos,´
G = (V, E ) be an interval graph having a chordless cycle [ v _{0} ,v _{1} ,v _{2} ,
with > 3, and let I _{k} denote the interval corresponding to vertex v _{k} . Choose a
Interval graphs are chordal. As a matter of fact, Let
1957)
−1 ,v 0 ]
,v
, points p _{i} constitute a strictly increasing or strictly decreasing sequence. Therefore, it is impossible for I _{0} and I _{} _{−}_{1} to intersect, contradicting the assumption that {v _{0} ,v _{} _{−}_{1} } ∈ E.
point p _{i} ∈ I _{i}_{−}_{1} ∩I _{i} , for i = 1, 2,
− 1. Since I _{i}_{−}_{1} and I _{i} _{+}_{1} do not overlap, the
Example
interval model, and is chordal.
K _{n} has a permutation model, a circle model, a circulararc model, an
Technical University of Catalonia
29/152
Gabriel Valiente
Student Presentations
• Prove the claimed relationships between the ﬁve classes of intersection graphs (chordal, circle, circulararc, interval, and permutation graphs).
• WenLian Hsu, Ross M. McConnell. PC Trees and CircularOnes Arrangements. Theoretical Computing Science, 296(1):99–116, 2003.
• WenLian Hsu. A Simple Test for the Consecutive Ones Property. Journal of Algorithms, 43(1):1–16, 2002.
• Emmanuel Gasse. A Proof of a Circle Graph Characterization. Discrete Mathematics, 173(1–3):277283, 1997.
• Stephan Olariu. An Optimal Greedy Heuristic to Color Interval Graphs. Information Processing Letters, 37(1):21–25, 1991.
• WeiKuan Shih, T. C. Chern, WenLian Hsu. An O ( n ^{2} log n ) Algorithm for the Hamiltonian Cycle Problem on CircularArc Graphs. SIAM Journal on Computing, 21(6):1026–1046, 1992.
Technical University of Catalonia
30/152
Gabriel Valiente
Algorithms on Permutation Graphs
Technical University of Catalonia
31/152
Gabriel Valiente
Algorithms on Permutation Graphs
Deﬁnition
A permutation graph is a graph that is isomorphic to the intersection
graph of a ﬁnite set of chords between an ordered set of vertices and a permutation of them.
Example
The following ﬁnite set of chords between an ordered set of vertices
and a permutation of them is a permutation model or matching diagram of the permutation graph shown to the left.
Remark
e _{6}
e
e
8
1
e 2
e _{7}
e
4
e 5
e
3
v 1 
v 2 
v 3 
v 4 
v 5 
v 6 
v 7 
v 8 

v 3 
v 5 
v 8 
v 4 
v 6 
v 1 
v 7 
v 2 
A permutation graph is the graph of inversions in a permutation.
Technical University of Catalonia
32/152
Gabriel Valiente
Algorithms on Permutation Graphs
• Recognition of Permutation Graphs
Paul C. Gilmore, A. J. Hoffman. A Characterization of
Comparability Graphs and of Interval Graphs. Canadian Journal of Mathematics, 16(3):539–548, 1964.
Shimon Even, Amir Pnueli, Abraham Lempel. Permutation Graphs
and Transitive Graphs. Journal of the ACM, 19(3):400–410, 1972.
Martin C. Golumbic. The Complexity of Comparability Graph
Recognition and Coloring. Computing, 18(4): 199–208, 1977.
Jeremy Spinrad. On Comparability and Permutation Graphs. SIAM Journal on Computing, 14(3):658–670, 1985.
O
( n + m )
Dieter Kratsch, Ross M. McConnell, Kurt Mehlhorn, Jeremy P.
Spinrad. Certifying Algorithms for Recognizing Interval Graphs and
Permutation Graphs. Proc. 14th Annual ACMSIAM Symp. Discrete Algorithms, 2003, pages 158–167.
O
( n
^{3}
)
O
( n
^{3}
)
O
( n
^{3}
)
O
( n
^{2}
)
Technical University of Catalonia
33/152
Gabriel Valiente
Algorithms on Permutation Graphs
• Maximum Independent Set of Permutation Graphs
O 
( n log n ) 
Haklin Kim. Finding a Maximum Independent Set in a 
Permutation Graph. Information Processing Letters, 36(1):19–23, 

1990. 

O 
( n log n ) 
D. T. Lee, Majid Sarrafzadeh. Maximum Independent Set of a 
Permutation Graph in k Tracks. International Journal of Computational Geometry and Applications, 3(3):291–304, 1993. 

O 
( n log n ) 
Peter Widmayer, C. K. Wong. An Optimal Algorithm for the 
Maximum Alignment of Terminals. Information Processing Letters, 20(2):75–82, 1985.
Technical University of Catalonia
34/152
Gabriel Valiente
Algorithms on Permutation Graphs
• Maximum Independent Set of Permutation Graphs
O 
( n loglog n ) 
MawShang Chang, FuHsing Wang. Efﬁcient Algorithms 
for the Maximum Weight Clique and Maximum Weight Independent Set Problems on Permutation Graphs. Information Processing Letters, 43(6):293–295, 1992. 

O 
( n loglog n ) 
Erkki Mäkinen. On the Longest Subsequence Problem for 
Permutations. International Journal of Computer Mathematics, 77(1):45–53, 2001. 

O 
( n loglog n ) 
Frederico Malucelli, Thomas Ottmann, Daniele Pretolani. 
Efﬁcient Labelling Algorithms for the Maximum Noncrossing 

Matching Problem. Discrete Applied Mathematics, 47(2):175–179, 

1993. 

O 
( n loglog n ) 
MingShing Yu, Lin Yu Tseng, ShoeJane Chang. 
Sequential and Parallel Algorithms for the MaximumWeight Independent Set Problem on Permutation Graphs. Information
Processing Letters, 46(1):7–11, 1993.
Technical University of Catalonia
35/152
Gabriel Valiente
Algorithms on Permutation Graphs
• Isomorphism of Permutation Graphs
O
( n ^{2} )
Charles J. Colbourn. On Testing Isomorphism of Permutation Graphs. Networks, 11(1):13–21, 1981.
Technical University of Catalonia
36/152
Gabriel Valiente
Algorithms on Permutation Graphs
Lemma (Gries, 1981) solved in O ( n log n ) time.
Proof
minimum values
The longest nondecreasing sequence problem can be
Scan the sequence ( v _{1} ,v _{2} ,
m _{1} ,m _{2} ,
,m
_{n} ) from left to right and maintain the
,v
_{k} which end the nondecreasing sequences of length
1, 2,
,
k found so far. Then, k is length of a longest nondecreasing sequence.
Each value v _{i} in the sequence is compared with the minimum value m _{1} of the shortest nondecreasing sequence and with the minimum value m _{k} of the longest nondecreasing sequence.
• If v _{i} < m _{1} , then m _{1} is set to v _{i} .
• If v _{i} m _{k} , then a nondecreasing sequence of length k + 1 is found, and m _{k} _{+}_{1} is set to v _{i} .
• If m _{1} v _{i} < m _{k} , then an index j is found such that m _{j}_{−}_{1} v _{i} < m _{j} , and m _{j} is set to v _{i} .
Finding the index j takes O ( log k ) time, yielding the O ( n log n ) time bound.
Technical University of Catalonia
37/152
Gabriel Valiente
Algorithms on Permutation Graphs
Lemma (Golumbic, 1980)
Let G be the (permutation) graph of inversions of a
permutation π. Then, the decreasing sequences of π are in onetoone correspondence with the cliques of G, and the increasing sequences of π are in onetoone correspondence with the independent sets of G.
Proof
adjacent if and only if the corresponding chords in a matching diagram of G intersect if and only if v _{i} and v _{j} are inverted in the permutation π.
Follows from two vertices v _{i} and v _{j} in a permutation graph G being
_{1} 
··· 
_{i} 
··· 
_{j} 
··· 
_{k} 
··· 
n 

i 

j 
k 

π 
··· 
j 
··· 
_{k} 
··· 
_{i} 
··· 
Technical University of Catalonia
38/152
Gabriel Valiente
Algorithms on Permutation Graphs
Lemma
graphs in O ( n loglog n ) time.
Proof
restricted universe being solvable in O ( n loglog n ) time.
The maximum independent set problem can be solved on permutation
Follows from the problem of ﬁnding the successor of an element in a
Technical University of Catalonia
39/152
Gabriel Valiente
Algorithms on Circle Graphs
Technical University of Catalonia
40/152
Gabriel Valiente
Algorithms on Circle Graphs
• Recognition of Circle Graphs
Walid Naji. Reconnaissance des Graphes de Cordes. Discrete Mathematics, 54(3):329–337, 1985.
André Bouchet. Reducing Prime Graphs and Recognizing Circle Graphs. Combinatorica, 7(3):243–254, 1987.
O ( mn )
Csaba P. Gabor, Kenneth J. Supowit, WenLian Hsu. Recognizing
Circle Graphs in Polynomial Time. Journal of the ACM, 36(3):435–473, 1989.
WenLian Hsu. O ( mn ) Algorithms for the Recognition and
Isomorphism Problems on CircularArc Graphs. SIAM Journal on Computing, 24(3):411–439, 1995.
Jeremy Spinrad. Recognition of Circle Graphs. Journal of Algorithms, 16(2):264–282, 1994.
O ( mn )
O ( n ^{2} )
O ( n
^{7}
)
O ( n
^{5}
)
Technical University of Catalonia
41/152
Gabriel Valiente
Algorithms on Circle Graphs
• Maximum Independent Set of Circle Graphs
O 
( n ^{3} ) Fanica Gavril. Algorithms for a Maximum Clique and a Maximum Independent Set of a Circle Graph. Networks, 3(3):261–273, 1973. 
O 
( n ^{3} ) Robin Liu, Simeon C. Ntafos. On Decomposing Polygons into Uniformly Monotone Parts. Information Processing Letters, 27(2):85–88, 1988. 
Technical University of Catalonia
42/152
Gabriel Valiente
Algorithms on Circle Graphs
• Maximum Independent Set of Circle Graphs
O 
( n ^{2} ) Takao Asano, Tetsuo Asano, Hiroshi Imai. Partitioning a Polygonal Region into Trapezoids. Journal of the ACM, 33(2):290–312, 1986. 
O 
( n ^{2} ) Takao Asano, Hiroshi Imai, Akira Mukaiyama. Finding a Maximum Weight Independent Set of a Circle Graph. IEICE Transactions, E74(4):681–683, 1991. 
O 
( n ^{2} ) Olivier Goldschmidt, Alexan Takvorian. An Efﬁcient Algorithm for Finding a Maximum Weight Independent Set of a Circle Graph. IEICE Transactions, E77A(10):1672–1674, 1994. 
O 
( n ^{2} ) Ronald C. Read, Doron Rotem, Jorge Urrutia. Orientations of Circle Graphs. Journal of Graph Theory, 6(3):325–341, 1982. 
O 
( n ^{2} ) Kenneth J. Supowit. Finding a Maximum Planar Subset of a Set of Nets in a Channel. IEEE Transactions on ComputerAided Design, 6(1):93–94, 1987. 
Technical University of Catalonia
43/152
Gabriel Valiente
Algorithms on Circle Graphs
• Maximum Independent Set of Circle Graphs
O
( nd )
Alberto Apostolico, Mikhail J. Atallah, Susanne E. Hambrusch. New Clique and Independent Set Algorithms for Circle Graphs.
Discrete Applied Mathematics, 36(1):1–24, 1992; 41(2):179–180,
1993.
Technical University of Catalonia
44/152
Gabriel Valiente
Algorithms on Circle Graphs
• Isomorphism of Circle Graphs
O ( mn )
WenLian Hsu. O ( mn ) Algorithms for the Recognition and
Isomorphism Problems on CircularArc Graphs. SIAM Journal on
Computing, 24(3):411–439, 1995.
Technical University of Catalonia
45/152
Gabriel Valiente
Algorithms on Circle Graphs
Deﬁnition
of a ﬁnite set of chords of a circle.
Example
circle graph shown to the left.
A circle graph is a graph that is isomorphic to the intersection graph
The following ﬁnite set of chords of a circle is a chord model of the
e 3
e 1
e 4
e 5
e 2
e _{6}
e _{7}
v
v
8
v
v
5
v
4
7
9
v 6
v 10
v 3
v
v
v 13
2
v
14
v 11
v 12
1
Technical University of Catalonia
46/152
Gabriel Valiente
Algorithms on Circle Graphs
circle graph
nonadjacency
adjacency
disjointness
containment
overlap
interval model
• A set of mutually disjoint, or contained in each other, intervals of I ( G) models an independent set of G.
• A set of mutually overlapping intervals of I ( G ) models a clique of G.
Technical University of Catalonia
47/152
Gabriel Valiente
Algorithms on Circle Graphs
( 3, 6 ) cage
P. J. Heawood (1861–1955)
Technical University of Catalonia
48/152
Gabriel Valiente
Algorithms on Circle Graphs
( 3, 7 ) cage
W. F. McGee (1937–)
Technical University of Catalonia
49/152
Gabriel Valiente
Algorithms on Circle Graphs
( 3, 8 ) cage
W. T. Tutte (1917–2002) and H. S. Coxeter (1907–)
Technical University of Catalonia
50/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 1
r = 1
Technical University of Catalonia
51/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 2
r = 2
Technical University of Catalonia
52/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 3
r = 4
Technical University of Catalonia
53/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 4
r = 8
Technical University of Catalonia
54/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n 
12345 
r 
1248 
Technical University of Catalonia
55/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 5
r = 16
Technical University of Catalonia
56/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n 
123456 

r 
1 
2 
4 
8 
16 
Technical University of Catalonia
57/152
Gabriel Valiente
Algorithms on Circle Graphs
Exercise
by joining n points on the circle with chords
Maximum number of nonoverlapping regions r that can be obtained
n = 6
r = 31
Technical University of Catalonia
58/152
Gabriel Valiente
Algorithms on Circle Graphs
Deﬁnition
chords in the chord model of the graph
The total chord length of a circle graph is the sum of the length of all
e 3
e 1
e 4
e 5
e 2
e _{6}
e _{7}
v
v
8
v
v
5
v
4
7
9
v 6
v 10
v 3
v
v
v 13
2
v
14
v 11
v 12
1
Example
The total chord length of the previous circle graph is 37
Technical University of Catalonia
59/152
Gabriel Valiente
Algorithms on Circle Graphs
In the interval model of a circle graph, the total order on the vertex set along the circumference of a circle is replaced by a total order along the line
Technical University of Catalonia
60/152
Gabriel Valiente
Algorithms on Circle Graphs
Deﬁnition
The density of a circle graph is the maximum number of intervals
Example
The density of the previous circle graph is 4
Technical University of Catalonia
61/152
Gabriel Valiente
Algorithms on Circle Graphs
• An interval model of a circle graph can be obtained from the chord model of the circle graph by a simple transformation, consisting in cutting the circumference of the circle at some point p which is not an endpoint of a chord and unfolding it at point p
• The chord model can be reconstructed by wrapping around the circle the collection of intervals on the line
• Given a chord model of a circle graph, for each choice of point p in the previous transformation a different interval model is obtained, and both the density and the total chord length of the circle graph depend on the particular chord or interval model chosen, that is, on the choice of point p for the transformation between the chord model and the interval model of the circle graph
Technical University of Catalonia
62/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 37 and density 4
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
63/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 43 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
64/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 37 and density 4
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
65/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 43 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
66/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 47 and density 6
7
8
9
6
5 4
3
^{1}^{0} 11
12 ^{1}^{3}
Technical University of Catalonia
67/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 41 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
68/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 45 and density 6
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
69/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 39 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
70/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 41 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
71/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 37 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
72/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 45 and density 6
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
73/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 41 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
74/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 47 and density 6
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
75/152
Gabriel Valiente
Algorithms on Circle Graphs
Example
The following chord model and corresponding interval model for the
previous circle graph have total chord length 39 and density 5
7
8
9
6
5
^{1}^{0} 11
4
3
12 ^{1}^{3}
Technical University of Catalonia
76/152
Gabriel Valiente
Algorithms on Circle Graphs
It will be assumed, without loss of generality, that all chords are ordered from lower to higher vertex number, that is, i < j for all chords ( v _{i} ,v _{j} ) ∈ E in a circle
graph
G = (V, E )
• The length of chord e = (v _{i} ,v _{j} ) ∈ E is len ( e ) = j − i
• MIS ( G ) denotes a maximum independent set of G
• For all chords ( v _{i} ,v _{j} ) = e ∈ E, MIS ( e ) denotes a maximum independent set of the subgraph of G induced by {v _{k} ∈ V 
Molto più che documenti.
Scopri tutto ciò che Scribd ha da offrire, inclusi libri e audiolibri dei maggiori editori.
Annulla in qualsiasi momento.